Live from GTC: Write Once and Run Anywhere with NVIDIA's cuTile

This episode dives into one of the biggest challenges in AI development: seamlessly moving from local experimentation to large-scale deployment. Host Logan Lawler is joined by NVIDIA AI and Cloud Systems Software Engineers Rohit Pujar-Nagraj and Saish Komalla to explore how new tools like cuTile are simplifying that journey. They break down how developers can write code once on a desktop system and deploy it across powerful data center GPUs with minimal friction. The conversation highlights how hardware-agnostic development is unlocking faster workflows and near-linear performance scaling across environments. It’s a quick, practical look at how AI teams can go from desk-side development to cloud or data center production, without rewriting everything along the way. You can also watch previous episodes here. Follow Us LinkedIn @DellTechnologies Twitter @DellTech YouTube @DellTechnologies LinkedIn @NVIDIA Twitter @NVIDIA Instagram @NVIDIA Facebook @NVIDIA Presented by Dell and NVIDIAhttps://www.dell.com/precisionai https://www.dell.com/dellpromax https://www.dell.com/nvidia-ai Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.